The Prostate Cancer Prevention Trial (PCPT) demonstrated that finasteride reduces the prevalence of prostate cancer by 24.8% (risk reduction) but questions remain regarding the cost-effectiveness of widespread utilization. The purpose of the current analysis was to evaluate the cost-effectiveness of chemoprevention utilizing a quality-of-life adjustment.
A Markov decision analysis model with probabilistic sensitivity analysis was designed to determine the lifetime prostate health-related costs, beginning at age 50 years, for men treated with finasteride compared with placebo. Model assumptions were based on data from the PCPT; Surveillance, Epidemiology, and End-Results program; literature review of costs, utilities, and transition rates among various prostate cancer health states; and local institutional cost data.
The quality-adjusted cost-effectiveness ratio for finasteride compared with men not receiving chemoprevention was $122,747 (in U.S.$) per quality-adjusted life-years saved (QALYs). If finasteride is assumed to not increase the incidence of high-grade tumors, the quality-adjusted cost-effectiveness ratio was $112,062 per QALYs. Sensitivity analysis found that chemoprevention of prostate cancer with an agent that has no effect on the prevalence of benign prostatic hyperplasia can render a cost-effectiveness ratio of <$50,000 per QALYs saved when applied to a high-risk population associated with a 25% risk reduction, and a cost of $30 per month.
Finasteride has been shown to be effective as a prostate cancer chemopreventive therapy in men.1 In the Prostate Cancer Prevention Trial (PCPT), the 7-year prevalence of prostate cancer was reduced by 24.8% in the finasteride versus placebo group. The analysis found a higher prevalence of higher-grade tumors among the finasteride-treated group, but there is ongoing debate regarding whether this effect is real or artifactual. Recent observations from the PCPT support the possibility that the disparity in grade distribution among the study populations may be a result of the improved sensitivity of prostate-specific antigen (PSA) and digital rectal examination (DRE) for detecting high-grade disease in the finasteride group.2, 3 The widespread implementation of chemoprevention depends on demonstrating a reduction in the risk of developing prostate cancer, an acceptable morbidity/toxicity for the intervention, and cost-effectiveness. Because of the long-term survival associated with prostate cancer, the PCPT was not designed to evaluate the survival benefits of finasteride. Several models have demonstrated a modest but meaningful survival benefit using finasteride.4, 5 The decision concerning the institution of a widespread prevention strategy depends not only on demonstrating a survival benefit, but also on utilizing an intervention that has minimal morbidity. Finasteride is a relatively nontoxic drug whose main morbidity is a reduction in libido, erectile function, and ejaculate volume in approximately 5% of patients each but with the benefit of reduction in lower urinary tract symptoms (LUTS) associated with benign prostatic hyperplasia (BPH).1, 6
We have previously published an analysis of the cost-effectiveness of chemoprevention using finasteride based on the PCPT.7, 8 This analysis found that finasteride use for chemoprevention was expensive, with a cost of $578,400 to $1107,000 (in U.S.$) per life-year (LY) saved. Widespread chemoprevention would be well above the widely accepted threshold for cost-effectiveness ($50,000-$100,000 per LY saved) unless there was a significant decrease in medication cost or finasteride was only used in populations with a very high incidence of cancer. However, the prevention of prostate cancer may have a significant impact on the quality of life for men by averting complications and deleterious effects associated with treatment. Recently, a comprehensive assessment of quality of life for men at various stages of prostate cancer progression was developed.9 This analysis quantified the relative importance of various health outcomes by generating mean utilities for multiple prostate cancer health states. The current analysis aims to incorporate health utilities for various states of prostate cancer to generate a cost per quality-adjusted life-years (QALYs) rendered by chemoprevention of prostate cancer.
MATERIALS AND METHODS
The study design is a cost-effectiveness analysis from a societal perspective. Similar to the PCPT entry criteria, the base case was assumed to have a PSA of ≤3.0 mg per mL and a normal DRE.1 The base case is a 50-year-old man who receives daily finasteride treatment for a total of 25 years or until prostate cancer is diagnosed. Incidence rates for prostate cancer were based on for-cause biopsy detection rates (and excluded cancer detected in end-of-study biopsies, which were not for cause) in the placebo arm of the PCPT. To estimate the age-specific incidence rates for men not represented in the PCPT (those aged <55 years), a smooth isotonic regression curve was applied to the age-specific detection rates observed in the PCPT. Incidence rates for men aged ≥75 years were based on Surveillance, Epidemiology, and End-Results (SEER) data.10 Incidence rates among finasteride-treated men were derived from the placebo-arm detection rates after applying a yearly relative risk reduction afforded by finasteride.1 Reduced incidence of cancer as a result of finasteride did not persist after discontinuation of the medication (aged ≥75). Men who experience adverse side effects, such as erectile dysfunction, loss of libido, and gynecomastia, as a result of chemoprevention are assumed to discontinue medication within the first year of initiating finasteride treatment. The noncompliant rate as a result of adverse side effects was based on data from the PCPT and was estimated at 18.3%.1 Men with symptomatic BPH were treated with medication (95%) or surgery (5%). Therefore, for men with symptomatic BPH who were receiving finasteride chemoprevention, the extra cost for medication treatment of BPH symptoms was equal to the cost of an alpha-blocker. For men not receiving chemoprevention, the cost of medication included alpha‒blocker therapy (60%), finasteride (30%), or a combination of alpha‒blocker therapy and finasteride (10%).
The cost per LY saved is the difference between the lifetime costs for the chemoprevention group and placebo group divided by the gain of LYs. Results for quality adjustment are displayed as cost per QALYs saved. An annual discount rate of 3% was applied to future costs and future years of life.11, 12 Discounting is necessary when the experience of the patient in the near term is valued more than future costs and health outcomes.11, 12 Model assumptions are shown in Tables 1 and 2. All costs were updated to 2007 U.S. dollars with the Gross Domestic Product Deflator Inflation Calculator.13 Markov modeling was designed with Treeage Pro Healthcare.14
Table 1. Model Assumptions for Health State Utilities (Mean and SD)
Table 2. Assumptions on Model Parameters Including Costs of Medication, Disease Status, Treatment Procedures, Visits, and Tests With Additional Assumptions on the Prevalence and Effectiveness of Finasteride on Benign Prostate Hypertrophy
The long-term progression and survival rates for patients treated with radical prostatectomy are based on literature regarding the natural history of treated prostate cancer.15, 16 To approximate the outcomes of men from the PCPT, we utilized publications characterizing outcomes among patients with low-stage disease (T1c and T2a-b) in the PSA era when available. Transition rates for biochemical recurrence at 5 years, 10 years, and 15 years from treatment were based on validated nomograms for various cancer grades.15 Transition rates for metastasis and death were based on data by Pound et al.16
Health State Utilities
Estimates of quality of life for various prostate cancer- and treatment-related health states are shown in Table 1. These were adapted from prior studies examining outcomes in men with and without prostate cancer.9, 17, 18 In these reports, values are assigned to various health states on a scale ranging from 0 (dead) to 1 (perfect health). Disutility for men with symptomatic BPH was estimated to be 0.05 based on a higher disutility of 0.12 in men with untreated BPH9 because we assumed that all men with symptomatic BPH were treated. For postprostatectomy NED (no evidence of disease), asymptomatic metastasis, and symptomatic metastasis health states, a standard deviation of 0.06 was applied the group but the relative impact of worse disease states on quality of life (QOL) were maintained.
Reduction in the Prevalence of BPH
The influence of finasteride on the reduction of LUT symptoms attributable to BPH was included in the model.19 However, for this analysis to be applicable to other potential chemoprevention agents, an additional model was developed that omitted all variables related to BPH (cost of treatment for BPH, QOL disutility from BPH, and the reduction in the prevalence of BPH afforded by chemoprevention).
Sensitivity analysis was performed to investigate the effect of adjusting several of the base case assumptions. We performed sensitivity analyses after adjustment of the cost and risk reduction provided by chemoprevention. In addition, chemoprevention for high-risk populations such as men with a family history of prostate and/or men with high-grade prostatic intraepithelial neoplasia may be more cost-effective because of the increased prevalence of the disease. We examined the effects of adjusting the prevalence of prostate cancer in the population. Uncertainties in values are considered simultaneously using probabilistic sensitivity analysis. The probability distributions for model variables were derived from the literature.9 For cost estimates, baseline values ± 25% were used to define the triangular distributions; 5000 Monte-Carlo samples were used for probabilistic sensitivity analysis.
Reduced Overall Prostate Cancer Incidence, Increased High-Grade Prostate Cancer With Finasteride
Our first analysis assumed the grade distribution of the PCPT-reduced overall prostate cancer and increased high-grade prostate cancer in the finasteride versus the placebo arm.1 Under this assumption, the mean (standard deviation [SD]) QALYs after age 50 years for men treated with finasteride was 17.587 (0.156) and this chemoprevention strategy cost $14,470 (1085) per person. This compared with a mean (SD) QALYs of 17.513 (0.209) and a cost of $5445 (237) for men not receiving finasteride. In this analysis, finasteride chemoprevention is associated with a gain of 74 QALYs per 1000 men and the quality-adjusted cost-effectiveness ratio for finasteride compared with men not receiving chemoprevention was $122,747 per QALYs.
Reduced Overall Prostate Cancer Incidence, No Increased High-Grade Prostate Cancer With Finasteride
Based on suggestions that the higher distribution of Gleason score ≥7 in the finasteride versus the placebo arm of the PCPT may have been more apparent than real,2, 3 our second analysis assumed that finasteride reduces overall prostate cancer and does not increase the incidence of high-grade disease. Under this assumption, the mean (SD) QALYs after age 50 years for men receiving finasteride was 17.593 (0.161) and this chemoprevention strategy cost $14,484 (1081) per person. In this analysis, finasteride chemoprevention is associated with a gain of 80 QALYs per 1000 men and the discounted quality-adjusted cost-effectiveness ratio for finasteride compared with men not receiving chemoprevention was $112,062 per QALYs.
Model Without BPH-Related Variables
The beneficial effect of finasteride on the prevalence of BPH contributes a substantial portion to the perceived benefit of finasteride chemoprevention because of the large risk reduction (40%) in BPH prevalence by finasteride. An additional analysis that omits this beneficial effect and the cost and disutility from BPH was performed to evaluate only the effects of finasteride as a cancer prevention agent. The mean (SD) QALYs after age 50 years for men in the chemoprevention arm was 17.690 (0.048) and this chemoprevention strategy cost $153,216 (1061) per person. This compared with a mean (SD) QALYs of 17.646 (0.057) and a cost of $3355 (202) for men not receiving finasteride. Finasteride chemoprevention without beneficial effect on prevalence of BPH is associated with a gain of 44 QALYs per 1000 men and the discounted quality-adjusted cost-effectiveness ratio for finasteride (assuming no impact on BPH) compared with men not receiving chemoprevention was $224,683 per QALYs.
A 2-way sensitivity analysis relaxing the assumptions regarding the disease prevalence and the relative risk reduction in the incidence of prostate cancer is shown in Tables 3 and 4. This analysis assumes a similar grade distribution among the chemoprevention and placebo arms. The cost-effectiveness ratios decrease as the prevalence of prostate cancer or the relative risk reduction afforded by chemoprevention increase. Based on prostate cancer incidence rates observed in the PCPT (men aged ≥55 years) and SEER data (for men aged ≤75 years), the model assumes a prevalence of prostate cancer for men aged ≥50 years of 14.6%. When applied to a population at higher risk of developing prostate cancer (prevalence ≥30% for men aged ≥50 years), substantial improvement in cost-effectiveness is achieved. This table also illustrates the relative contribution to the cost-effectiveness of finasteride's effect to reduce the prevalence of BPH.
Table 3. Incremental Cost-Effectiveness of Chemoprevention by High-Risk Subgroup and Relative Risk Reduction*
U.S.$ per LYs saved
U.S.$ per QALYs saved (No drug effect on prevalence of BPH)
U.S.$ per QALYs saved (Medication decreases prevalence of BPH)
Table 4. Incremental Cost-Effectiveness of Chemoprevention, by High-Risk Subgroup, Relative Risk Reduction, and Cost of Chemoprevention*
Relative risk reduction
Cost of chemoprevention, U.S.$
Lifetime prevalence (Men diagnosed after age 50 years)
Assuming a similar grade distribution between the chemoprevention and placebo arms and no reduction in the prevalence of benign prostatic hyperplasia. Results are cost per quality-adjusted life-years saved.
Graphic representations of sensitivity analyses evaluating the cost per QALYs saved as determined by the cost of chemoprevention at various prevalence rates are shown in Figure 1. Because of the finding that 5-alpha reductase inhibitors have the additional impact on QOL of reducing BPH-related symptoms, we evaluated the impact of varying cost of a drug at various prevalence rates on cost of chemoprevention when there is no effect on BPH (Fig. 2). Targeting high-risk populations (prevalence ≥30% for men aged ≥50 years) consistently renders a cost/effectiveness < $100,000 as long as the cost of the chemoprevention is < $1000 per year.
Figure 3 shows a graphic representation of incremental cost-effectiveness as a function of model termination in years for both the LYs and QALYs saved models. Benefit from chemoprevention is appreciated earlier in the quality-adjusted model compared with the LY saved model. Because death from prostate cancer is a late and infrequent event in most cases, QOL benefits are demonstrated well before survival benefits are noticeable.
The results of the current analysis indicate that chemoprevention with finasteride is unlikely to be cost-effective when considering only the impact on survival differences among treated versus untreated groups. However, quality adjustment analysis reveals a significant improvement in the cost-benefit ratio and even approaches a willingness-to-pay threshold of $100,000 per QALYs.
Benefits in the LY saved models are earned when there is a difference in mortality as a result of the intervention. In our model, finasteride for chemoprevention has a limited impact on prostate cancer mortality. As a result, incremental cost-effective values for the LYs saved model are much higher than the willingness-to-pay threshold of $100,000. This is consistent with previous analyses that have demonstrated a modest impact of chemoprevention with finasteride on survival.5, 7, 8, 20
Although most men who are diagnosed with prostate cancer in the modern era do not die of prostate cancer, they do experience considerable morbidity and adverse effects on QOL from this disease. Estimates of the incidence of erectile dysfunction and urinary incontinence have been as high as 70% to 91% and 39% to 52%, respectively, after treatment with primary disease.20, 21 Furthermore, biochemical disease recurrence after primary treatment of prostate cancer is common, with up to 39% of men found to have a rising PSA at 15 years after treatment for clinically localized disease.22 The life expectancy for men with biochemical recurrence is good, but their QOL may be significantly reduced because of resulting treatments and anxiety.9 Through modeling with quality adjustment, benefits from chemoprevention can be accrued throughout a person's lifetime. As shown in Figure 3, there is a steeper decline in the incremental cost-effectiveness for the LYs model compared with the QALYs model because the benefits of finasteride, with regard to mortality, are not appreciated until later in life.
A recently published cost-effectiveness model for chemoprevention with finasteride by Zeliadt et al20 found finasteride to offer a considerable cost burden with modest survival benefit at an incremental cost-effectiveness of $1.66 million per LY saved. This result is similar to our findings for cost per LYS (incremental cost-effectiveness >$1 million per LYS).7, 8 However, our quality-adjusted analysis found chemoprevention of prostate cancer to be more cost-effective than the results from their study. In their analysis, quality-adjustment revealed a cost of $200,000 per QALYs saved and a $100,000 per QALYs saved was not attained unless finasteride was reduced by 50% from its current cost. Our quality-adjusted analysis found chemoprevention with finasteride to slightly above $100,000 per QALYs saved even at its current price ($66 per month). This difference may be explained, at least in part, by their quality adjustment for men with prostate cancer. Unlike our analysis, which assigned a unique health state utility for various states of prostate cancer, they utilized an average disutility of 0.03 for all men with prostate cancer regardless of stage of disease. In addition, in their analysis prostate cancer incidence rates were based on SEER data. In contrast, we utilized prostate cancer incidence rates found in the PCPT based on men diagnosed for cause to estimate rates that might be found in a population more heavily screened than the SEER population. As a result, our model diagnosed more patients with prostate cancer and therefore more patients benefited from chemoprevention.
One of the benefits of our model is that it is applicable to other potential chemoprevention strategies. Using our sensitivity analyses, it is possible to modify the parameters to model the potential cost-effectiveness of other future interventions. The REDUCE trial and SELECT trial are 2 ongoing chemoprevention trials whose primary outcomes will be reduction in prostate cancer incidence. Models will be necessary to estimate benefits in survival and cost-effectiveness. Other considerations include changes in cost of medications. Although the current cost of generic finasteride is not significantly cheaper than the brand Proscar (Merck and Co. Inc, Whitehouse Station, NJ), future reductions in the cost of medication might improve the cost-effectiveness of this intervention. Furthermore, companies who are manufacturing new drugs may want to consider how to price their drugs to allow them to be cost-effective for widespread use.
The cost-effectiveness of a chemopreventive agent is dependent on the baseline risk of the population receiving it. Applying chemoprevention to a high-risk group such as men with high-grade prostatic intraepithelial neoplasia or men with a family history of prostate cancer may improve the cost-effectiveness ratio. Therefore, the sensitivity analyses in this model can be used to determine appropriate target populations for chemoprevention strategies. As our capacity to identify high-risk populations improves and as new chemopreventive agents are tested, it is quite possible that cost-effective agents will become available.
This model found chemoprevention with finasteride to increase the QALYs by 29 days per individual, which corresponds to a gain of 74 QALYs per 1000 men. Although seemingly small, these gains are not insignificant. In prevention strategies the differences in life expectancy are often no more than a few weeks or days because the gains are averaged across the entire target population.23 Indeed, Wright and Weinstein23 consider a gain of 1 month because of a preventative strategy aimed at an average risk population to signal a large difference. Examination of other prevention strategies puts this into perspective. Childhood immunizations for measles, rubella, and pertussis, for example, increase life expectancy by 0.1 month each.13 Cost-effective modeling for breast cancer chemoprevention found tamoxifen to increase survival outcomes by 42 days for women aged 50 years or 27 days for women aged 60 years.24
There are several limitations to this analysis. There is theoretical evidence that the higher percentage of high-grade cancers observed in the PCPT may be a result of sampling bias,25 but this issue is still unresolved. This analysis assumes that finasteride would have similar results in men that were not examined in the PCPT because of exclusion criteria (ie, age <55 years and significant LUTS). In addition, we do not know whether there are sustained effects of risk reduction in prostate cancer after discontinuation of the drug. Some inaccuracy of cost and transition rates is unavoidable given the variability of different practice patterns, local costs, and differences in prostate cancer outcome found in published series. Moreover, costs important from a societal perspective such as loss of productivity or travel were not considered.26 Finally, we utilized a willingness-to-pay threshold of $100,000 to declare cost-effectiveness, but this value is not well established. There is debate about the appropriate value for willingness-to-pay threshold, but $50,000 to $100,000 is often cited in the literature as the cost-effectiveness threshold. The $50,000 cutoff comes from an analysis performed for Medicare to address the cost-effectiveness of dialysis for patients with chronic renal failure.27 Because this figure was generated in 1982, critics have argued that this value has not accounted for inflation and may be too low.28 Ubel et al28 maintain that this value should change over time and they suggest that a value of $100,000 for current analyses may be too low. Laupacis et al29 propose a graded approach to adoption of interventions based on a willingness-to-pay threshold. To provide guidelines, they suggest that interventions that cost <$20,000 per QALY were considered to have strong evidence for adoption, and those costing >$100,000 per QALY were relatively cost-ineffective.
Despite these uncertainties with regard to the effects of finasteride and the assumptions made in the model, this analysis has important implications. It is clear that finasteride is unlikely to be cost-effective without consideration of the effect on QOL. In this analysis, quality-adjustment analysis reveals a significant improvement in the cost-effectiveness of finasteride for chemoprevention and even approaches a willingness-to-pay threshold of $100,000 per QALYs.